> Sorry, didn't mean to imply it wouldn't. You added this caveat though: > > > Note however that this really only does what you want when > tableB is initially empty
Yeah, that's fairly easy to work around though. Assuming you know the structure of the ultimate destination table, which seems fairly reasonable given that an INSERT statement is being constructed for that table, creating a temporary table with the same structure, inserting into that table and then moving the data from the temp table to its ultimate destination is a fairly simple matter. Seems so to me, anyway, but it is my idea, so... :) -Tom